(754c) Kinetics and Mechanism of Homogeneous Ice Nucleation in Freestanding Nanofilms of Supercooled Water

The question of how a vapor-liquid interface affects the kinetics and mechanism of ice nucleation at its vicinity has baffled the scientific community for decades [1], and is considered one of the ten biggest open questions about ice [2]. It is particularly challenging to address this question experimentally, considering the insufficient spatiotemporal resolution of the existing experimental techniques that are incapable of probing the emergence of nano-sized ice nuclei in the liquid. Therefore, it is only possible to address this question indirectly (e.g., nucleation rate measurements in droplets of different sizes), and even such indirect approaches also lack the resolution needed to conclusively address this question [3].

Not surprisingly, molecular simulations have emerged as attractive alternatives in the quest for solving this long-standing conundrum. Up until recently, however, direct calculations of homogeneous ice nucleation rate had only been achieved [4] for the coarse-grained mW model [5]. By comparing the rate of homogeneous ice nucleation in the bulk and in nanofilms and nanodroplets of supercooled water, it has been consistently demonstrated that in the case of mW, freezing is suppressed in the vicinity of a vapor-liquid interface [6-8]. For the more realistic molecular models of water, however, no such comparison was possible as direct calculations of rate were out of reach. Recently, we utilized [9] a coarse-grained variant of a path sampling technique, known as forward-flux sampling (FFS) [10], for computing the rate of homogeneous ice nucleation for the molecular TIP4P/Ice potential [11] . In this work, we utilize the same technique to compute volumetric rates of ice nucleation in freestanding nanofilms of supercooled water simulated using the TIP4P/Ice potential. We discuss the quantitative and qualitative differences between the TIP4P/Ice films and mW films in order to understand how the utilization of a molecular model with full electrostatic interactions might affect the kinetics and mechanism of ice nucleation.

1- A. Tabazadeh, Y. S. Djikaev, et al., Proc. Natl. Acad. Sci. USA, 99, 15873 (2002).

2- T. Bartels-Rausch, Nature, 494, 27 (2013).

3- O. F. Sigurbjornsson, R. Signorell, Phys. Rev. E, 77, 051601 (2008).

4- T. Li, D. Donadio, et al., Phys. Chem. Chem. Phys., 13, 19807 (2011).

5- V. Molinero, E. B. Moore, J. Phys. Chem. B, 113, 4008 (2009).

6- T. Li, D. Donadio, et al., Nat. Comm., 4, 1887 (2013).

7- A. Haji-Akbari, R. S. DeFever, et al., Phys. Chem. Chem. Phys.,16, 25916 (2014).

8- M. M. Gianetti, A. Haji-Akbari, et al., Phys. Chem. Chem. Phys., 18, 4102 (2016).

9- A. Haji-Akbari, P. G. Debenedetti, Proc. Natl. Acad. Sci. USA, 112, 10582 (2015).

10- R. J. Allen, D. Frenkel, et al., J. Chem. Phys., 124, 194111 (2006).

11- J. L. F. Abascal, E. Sanz, et al., J. Chem. Phys., 122, 234511 (2005).